Sedimentary and Foraminiferal Evidence of the 2011 Tōhoku-Oki Tsunami on the Sendai Coastal Plain, Japan

Sedimentary Geology 282 (2012) 78–89

Contents lists available at SciVerse ScienceDirect

Sedimentary Geology

journal homepage: www.elsevier.com/locate/sedgeo

Sedimentary and foraminiferal evidence of the 2011 Tōhoku-oki tsunami on the Sendai coastal plain, Japan

Jessica E. Pilarczyk a,b,⁎, Benjamin P. Horton a, Robert C. Witter c, Christopher H. Vane d, Catherine Chagué-Goff b,e, James Goff b a Sea Level Research, Department of Earth and Environmental Science, University of Pennsylvania, 240 S. 33rd Street, Philadelphia, PA 19104‐6316, USA b Australia-Paciﬁc Tsunami Research Centre, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney 2052, Australia c United States Geological Survey/Alaska Science Center, Anchorage, AK 99508, USA d British Geological Survey, Keyworth, Nottingham NG92BY, UK e Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, NSW 2232, Australia article info abstract

Article history: The 2011 Tōhoku-oki megathrust earthquake (Mw 9.0) generated a tsunami that reached the Sendai coastal Received 16 March 2012 plain with flow heights of ~2 to 11 m above TP (Tokyo Peil). We examined the tsunami deposit exposed in 14 Received in revised form 23 August 2012 shallow trenches along a ~4.5‐km transect perpendicular to the coast. We primarily document the strati- Accepted 27 August 2012 graphical, sedimentological, foraminiferal and geochemical characteristics of the Tōhoku-oki tsunami deposit Available online 7 September 2012 and perform a preliminary comparison with sediments deposited by the Jōgan tsunami of A.D. 869. In the coastal forest and rice fields inundated by the Tōhoku-oki tsunami, a poorly sorted, dark brown soil is Keywords: Sendai Plain buried by a poorly sorted, brown, medium-grained sand deposit. In some trenches located more than 1.2 km 2011 Tohoku-oki tsunami inland, the sand is capped by a thin muddy-sand layer. The tsunami deposit, although highly variable in Jogan tsunami thickness, is generally thickest (25 cm) near the coastal dune and thins to less than 5 mm at ~4.5 km inland. Foraminifera The tsunami deposit was discriminated from the underlying soil by the appearance of recent and fossil foraminifera and a pronounced increase in grain size that fined upward and landward. The recent foraminifera preserved in the sandy facies of the deposit are rare and showed evidence of prolonged subaerial exposure (e.g. pitting, corrosion, fragmentation). Recent foraminifera likely originated from coastal dune and beach sediments that were breached by the tsunami. Calcified and sediment in-filled, fossil foraminifera are abundant and were eroded from sedimentary units and transported by fluvial or wave activity to Sendai Bay. Trends associated with test size (e.g. decreasing concentration of large test sizes with distance inland) are in agreement with grain size data. At two locations a decrease in total organic carbon and an increase in δ13C were found in the tsunami sand compared with the underlying soil, supporting a beach to intertidal origin for the upper unit. © 2012 Elsevier B.V. All rights reserved.

1. Introduction the Pacific northwest (Kelsey et al., 2005), North Sea (Bondevik et al., 2005), New Zealand (Goff et al., 2001), Kamchatka (Pinegina Records of past tsunamis developed from the sedimentary evi- et al., 2003) and the South Pacific(Goff et al., 2011), and in the dence they leave behind, improve our understanding of the frequency regions affected by the 1960 Chile (Cisternas et al., 2005) and 2004 of tsunamis by expanding the age range of events available for study Indian Ocean (e.g. Jankaew et al., 2008) earthquakes. (Morton et al., 2007). Proper hazard assessment depends on an The identification of tsunami deposits is often based on the recogni- awareness of tsunamis and their impacts on coastal geomorphology, tion of anomalous sand units in low-energy environments such as ecology and rapidly expanding coastal populations. Stratigraphical coastal ponds, lakes, and marshes, which can be supported by microfos- sequences of tsunami deposits are often used to estimate recurrence sil evidence. For example, the A.D. 1700 Cascadia tsunami can be iden- intervals and provide insight into their source (e.g. earthquakes, tified with confidence from a sand unit that tapers landward (often landslides, volcanic eruptions; Bernard and Robinson, 2009). Recon- for several kilometers), contains a mixed microfossil assemblage and structions have shown repeated tsunamis during the Holocene in coincides with stratigraphical evidence for abrupt coseimic subsidence (e.g. Hemphill-Haley, 1995; Hawkes et al., 2011). Foraminiferal taxonomy has been commonly used as an indicator of tsunami deposits (e.g. ⁎ Corresponding author at: Sea Level Research, Department of Earth and Environmental Mamo et al., 2009) and most taphonomic studies of foraminifera focus Science, University of Pennsylvania, 240 S. 33rd Street, Philadelphia, PA 19104‐6316, USA. Tel.: +1 215 573 5145. on time-averaging or lateral transport of tests with only semi- E-mail address: [email protected] (J.E. Pilarczyk). quantitative observations on test condition (e.g. Hawkes et al., 2007;

Kortekaas and Dawson, 2007; Uchida et al., 2010). Recent research has have not been studied. The tsunami sand deposit extended 2.8 km in- shown that test condition provides further information regarding ener- land, however, since the Sendai plain coastline has prograded 1 km gy regimes and transport history (e.g. Hawkes et al., 2007; Kortekaas over the last ~1000 years, the inland extent of the Jōgan sand deposit and Dawson, 2007; Uchida et al., 2010; Pilarczyk et al., 2011; Pilarczyk is now ~4 km inland (Sawai et al., 2008b; Sugawara et al., 2011b). and Reinhardt, 2012a, 2012b). The proxy toolkit to examine paleo-tsunamis has expanded fol- 2. Regional setting lowing the modern surveys on the 2004 Indian Ocean and 2009 South Pacific tsunamis (Chagué-Goff et al., 2011). While geochemis- The Sendai plain is a low-lying (less than 5 m above TP; Tokyo Peil; try has been used for more than two decades to provide evidence mean sea level in Tokyo Bay), wave-dominated, microtidal (mean tidal for marine inundation in paleo-tsunami studies (e.g. Minoura and range of ~1 m) coastal plain, which extends approximately 50 km on Nakaya, 1991; Minoura et al., 1994; Goff and Chagué-Goff, 1999; the Pacific coast of north-eastern Japan (Fig. 1a). The area is bounded Chagué-Goff et al., 2002; Chagué-Goff, 2010), it has also recently by hills to the north, west and south, and a steep continental shelf gra- been used in studies of modern tsunamis (e.g. Chagué-Goff et dient (Tamura and Masuda, 2004, 2005). Early Miocene to Holocene al., 2011; Chagué-Goff et al., 2012–this issue). In this study we de- marine and non-marine sediments dominate the coastal plain with scribe foraminiferal assemblages (taxa and taphonomic), general Middle Jurassic to Early Cretaceous marine sedimentary rock outcrop- sedimentology (grain size) and geochemistry (TOC and δ13C) as ping to the north at Oshika Peninsula, as well as to the south near indicators of the 11th March 2011 Tōhoku-oki tsunami deposit. Sōma (Fig. 1b; Geological Survey of Japan, 2009). The area is character- Foraminifera and sedimentological data from a trench containing ized by a series of prograding Late Holocene beach ridges and freshwa- the A.D. 869 Jōgan tsunami deposit are also presented. Readers are re- ter back marshes. Sugawara et al. (2012) describe several artificial ferred to Szczuciński et al. (2012–this volume) for a detailed analysis geomorphic features (e.g. breakwaters, coastal forest, Teizan canal, irri- of the sedimentological and microfossil variability within the gation channels, etc.) that likely affected the run-up associated with the Tōhoku-oki tsunami deposit along the transect. Tōhoku-oki tsunami but were not in place in Jōgan time. Freshwater marshes along the Sendai plain support rice cultivation, with many 1.1. 2011 Tōhoku-oki tsunami fields interspersed with low-density housing (Fig. 1c). Main sediment sources to the area include three rivers (the Abukuma, Natori and On 11th March 2011 a great megathrust earthquake (Mw 9.0) Nanakita rivers) that account for much of the continued seaward along the Japan Trench generated a tsunami that reached the Sendai progradation of the coastline since the Middle Holocene (Saito, 1991; plain on the northeastern coast of Honshu, Japan (Fig. 1a) at 14:46 Tamura and Masuda, 2005). (Japan Standard Time) with run-up heights of 10–40 m (Mori et al., The study area (near Sendai airport) consists of four key environ- 2011). The earthquake ruptured over a distance of ~400 km with up- ments: coastal; coastal forest; paved landscape; and rice fields. The wards of 5‐m vertical and 24 to 60‐m lateral displacement of the sea- coastal zone transitions from intertidal marine (0.8 m below TP) to floor (Ito et al., 2011; Sato et al., 2011). The low-lying configuration of beach (1.7 m above TP) to artificial dune (2.3 m above TP; Figs. 1c, the coastal plain made Sendai particularly susceptible to tsunami in- 2, and 3a) within a distance of ~0.50 km from the shoreline. The arti- undation that reached 4.8 km inland (Association of Japanese ficial dune is composed of allochthonous sediment that was brought Geographers, 2011; Chagué-Goff et al., 2012–this issue), and in to armor the coastline (Fig. 2a). We do not know the origin of the sustained flooding several months after the event was documented dune sediment, but note that it consists of a core of loose gray sand at several locations (Sugawara et al., 2011a; Chagué-Goff et al., overlain by a ~0.3‐m layer of compacted, yellow-brown sand with a 2012–this issue). silt/clay matrix. We speculate that the upper compacted layer may be dredged material from nearby canals. The coastal forest (Figs. 1c 1.2. A.D. 869 Jōgan tsunami and 2b), ~0.26 km from the coastline at an elevation of ~2.3 m above TP, consists of mature pine trees that were planted 300 years Predecessors of the Tōhoku-oki tsunami (e.g. A.D. 1611 Keichō, ago as a means of protecting rice fields from salt spray (Sugawara A.D. 1793 Kansei, A.D. 1896 Meiji Sanriku, A.D. 1933 Showa Sanriku, and Imamura, 2010). Immediately landward of the coastal forest is a A.D. 1978) are numerous (Minoura and Nakaya, 1991; Minoura et paved landscape that contains several canals and a former wetland al., 2001; Sawai et al., 2008a, 2008b). However, only the Jōgan tsuna- that was used for greenhouse horticulture and farming (0.10– mi in A.D. 869 is considered to approach the 2011 tsunami in terms of 1.20 km from coastline, ~1.0 m above TP), and rice fields (1.20– its magnitude, area of coastline impacted (Sendai and Sōma regions) 4.50 km from coastline, −0.4–1.2 m above TP). and extent of inundation (greater than 2 km inland). On 13th July A.D. 869 an offshore earthquake approximately 200 km from the Sen- 3. Methods dai coastal plain (Fig. 1a) resulted in a large tsunami that caused widespread flooding. Estimates of the magnitude of the earthquake 3.1. Sample collection of the Tōhoku-oki tsunami deposit and that generated the Jōgan tsunami, as well as flow depths and inunda- geomorphology of the survey sites tion distances have been investigated in an attempt to improve tsunami hazard assessments (e.g. Satake et al., 2008; Sugawara et al., In May 2011 we examined 14 trenches (Figs. 1c and 3a) 2011b). containing the Tōhoku-oki tsunami deposit and its underlying soil The Jōgan deposit is a landward-thinning sand unit of variable along a transect from the coastal forest (trenches 5–9; 0.26–0.44 km thickness (~2–20 cm) that extends over the Sendai and Sōma regions inland), paved landscape (trenches 12–24; 0.54–1.23 km inland) (Sawai et al., 2008a; Sugawara and Imamura, 2010). The deposit con- and rice fields (trenches 31–86; 1.62–4.50 km inland). A topographic sists of well-sorted medium sand intercalated with terrestrial survey using a Real Time Kinematic GPS (RTK-GPS) with elevations tied organic-rich mud (Minoura et al., 2001). Overlying the Jōgan deposit to the Tokyo Peil was conducted at each sample location (Goto et al., is a thin soil unit that is capped by the yellowish felsic Towada-a 2011). tephra emplaced by a volcanic eruption ~250 km north of Sendai in Trenches were sampled for sediment (0.2‐ to 10‐cm resolution) A.D. 870–934 (Yamada and Shoji, 1981; Minoura et al., 2001; and described in terms of deposit thickness and sedimentological Machida and Arai, 2003). Marine and brackish water diatoms have composition. We examined each of the 14 trenches for lateral been documented within the deposit (Minoura et al., 2001), although changes (average of all samples obtained in the tsunami deposit) as- marine diatoms were rare in the Sendai area. Foraminifera, however, sociated with increasing distance inland and selected six sections for 80 J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89

a) 130°E 140°

N Hokkaido

40°N SEA OF JAPAN earthquake trench epicenter Sendai tsunami detail in b) source Japan

Tokyo Honshu Japan trenchPACIFIC OCEAN

100 km

30° b)

Oshika Penninsula N

Sendai Sedimentary unit: Mid to Late Jurassic marine sedimentary rock Late Pleistocene to Holocene fan deposits Early to Middle Triassic marine mudstone Late Pleistocene to Holocene (M & NM seds) Early Cretaceous marine sedimentary rock Early to Middle Miocene (M & NM seds) Middle to Late Miocene (M & NM seds) Late Miocene to Pliocene (M & NM seds) Late Pleistocene terrace Late Pleistocene to Holocene levee deposits Late Pleis. to Holocene sand dune deposits M & NM seds...marine and non-marine sediments 10 km

c) 82 86 58 N 66 36 31 24 48 19 12 500 m 29 7 4 3 17 2 Sendai airport 9 5 1

Agricultural land Beach/artificial dune Low density housing Coastal forest Canal Trench location

Fig. 1. (a) Map of Sendai, Japan showing broad-scale tectonics. The inferred fault zone rupture segmentation (purple shaded area) and the epicenter (black square) of the Tōhoku-oki earthquake are indicated (after Koper et al., 2011) along with the estimated source region of the A.D. 869 Jōgan tsunami (gray dotted ellipse; after Minoura et al., 2001). Location of trench transect (inset) and Jōgan trench (white circle) are indicated. (b) Geologic map of the Sendai plain and surrounding areas showing broad-scale sedimentary units (adapted from: Geological Survey of Japan, 2009). (c) Detailed map of the study area showing site locations of Tōhoku-oki trenches and surface samples (sites 1–4). J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89 81

a) spanning the entire coastal zone (sites 1–4; −0.09 to 0.14 km inland) for comparison with tsunami sediments.

Pacific ocean 3.2. Grain size analysis

We conducted grain size analysis using a Beckman Coulter laser dif- fraction particle size analyzer on all bulk (e.g. no grain size fraction removed) surface and trench samples. Prior to analysis, organics were removed and samples were stirred as a moist paste to homogenize the sediment and disaggregated with sodium hexametaphosphate following the methods of Donato et al. (2009). Grain size values for all surface samples and 14 trench sections were converted to the Wentworth-Phi Scale, interpolated and gridded using a Triangular Ir- Coastal forest regular Network (TIN) algorithm according to Sambridge et al. (1995), Artificial dune and plotted as Particle Size Distributions (PSDs) in Geosoft Oasis TM. Grain size descriptions follow that of Blott and Pye (2001).Dominant

grain size values for the 10th (D10) and 90th (D90) percentiles were calculated. Reporting a range of D values takes into account the skewness b) of a sediment distribution curve (e.g. ﬁne tail) and can be a more suit- able descriptor of grain size in non-Gaussian distributions (Blott and Pye, 2001). We used a Camsizer to calculate particle sphericity ranging from 0 (highly angular) to 1 (perfectly spherical). All grain size results are listed in Appendix 1.

3.3. Foraminiferal analysis

We conducted foraminiferal analysis on all surface and trench samples following the methods of Horton and Edwards (2006) where approximately 5‐cm3 samples were sieved (>63 μm) and examined in a liquid medium. Taxonomy followed Loeblich and Tappan (1987) and Hayward et al. (2004), and where possible, we counted up to 300 specimens (Patterson and Fishbein, 1989). Since we found no agglutinated species, samples were then dried at 25 °C, sieved and recorded as hav- ing small (b250 μm) or large (>250 μm) test sizes. We categorized individual specimens as recent (white; late Holocene) or fossil (robust, c) sediment in-filled and calcified; Miocene–Pleistocene; after Pilarczyk et al., 2011; Plate 1). Fossil foraminifera are easily identified since they generally maintain their test structure even after wave agitation disag- Muddy-sand gregates them from their parent rock (Pilarczyk et al., 2011). However, a long residence time in the nearshore environment results in significant abrasion and obscuring of diagnostic test features (e.g. aperture, sand sheet perforations, umbo, etc.) required for proper species identification. 10 cm Total number of individuals (fossil and recent individuals combined), (Dark laminae)

TSUNAMI total recent, total fossil and percent large specimens were enumerated. OVERWASH All foraminiferal results are listed in Appendices 2 and 3.

3.4. Stable carbon isotope and carbon analysis

Basal soil For measurement of δ13C and total organic carbon (TOC) in trenches, we selected two trenches (5 and 31), which include the soil, overlying sand and muddy-sand. Sediment samples were treated with 5% HCl for 18 h, washed with deionised water, dried in an oven at 40 °C overnight and milled to a fine powder using a pestle and Fig. 2. (a) Remnants of the artificial dune after the Tōhoku-oki tsunami. The tsunami mortar. Plant samples were treated with 5% HCl for 2–3 h, washed breached the dune in several locations and flattened pine trees and coastal shrubs, ex- with deionised water, dried in an oven at 40 °C overnight and milled fl posing an underlying stabilizing net. (b) Tsunami- attened coastal forest planted to a fine powder using a freezer mill (Vane et al., 2010).13C/12C anal- ~300 years ago to shelter rice fields from wind and salt spray. Large pine trees were flattened in a shore-normal direction, and in some cases uprooted, as a result of the yses were performed on sediment samples by combustion in a tsunami. (c) Tōhoku-oki deposit at Trench 31 (see Fig. 1c) indicating a ~10‐cm‐thick Costech Elemental Analyser coupled on-line to an Optima dual-inlet sand unit with a muddy-sand overlying basal rice field soil. mass spectrometer. δ13C values were calculated to the VPDB scale using a within-run laboratory standard (cellulose, Sigma Chemical prod. no. C-6413) calibrated against NBS-19 and NBS-22. Organic car- detailed analysis of vertical changes of the tsunami deposit with bon values (TOC wt/wt) were analyzed on the same instrument. Rep- depth (e.g. trenches 5, 12, 31, 36, 48, 86). We used a node near the licate analysis indicated a precision of b0.1‰ (1 StD) for δ13C and shoreline (38°8′27.64″N, 140°56′43.99″E) to calculate the distance 0.1% TOC (wt/wt) measurements. All sediments reported for geo- of each surface sample and trench location relative to the marine chemistry were sampled over a 1‐cm increment and are plotted as source. In addition we collected surface samples (upper 1 cm) an average depth. With the exception of the pine roots which have 82 J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89

Coastal Paved Rice fields a %N of 0.6 and C/N of 85.3, the %N values were below the limit of de- zone CF landscape Surface tection (LOD) of ~0.1%, therefore C/N values were unavailable to dis- a) 3 samples Tsunami trenches tinguish local from imported organic matter. All δ13C and TOC results 4 2 3 5 are listed in Appendix 4. 17 19 9 82 86 1 7 Dune 58 66 Elevation 10 12 24 31 36 ō 0 48 3.5. Analysis of the A.D. 869 J gan tsunami deposit 2 m TP 6 (meters above TP) 1 2 (m TP)

b) Flow height -1 ō CF...Coastal forest While the J gan deposit has previously been documented (Minoura 150 et al., 2001; Sawai et al., 2008a; Goto et al., 2011), we took the opportu- 3 100 nity to log and sample a single trench containing evidence of the A.D. 869 Jōgan tsunami in a rice ﬁeld ~10 km north of the Tōhoku-oki tran- 50 per 1 cm sect (Fig. 1a; 38°13′56.96″N, 140°58′33.82″E) and ~1.5 km from the Concentration of fossil foraminifera 0 * present coastline. We also conducted foraminiferal and grain size anal- c) 3 150 *3 individuals per cm ysis as outlined in Sections 3.2 and 3.3 and their results can be found in 3 100 Appendices 1 and 3.

50 4. Results per 1 cm

Concentration of 0 d) recent foraminifera We conducted a field survey to the north of Sendai airport 80 (Fig. 1a) and found evidence of flow heights of 10 to 11 m above TP behind artificially emplaced dunes (Figs. 2 and 3a). Approximately 40 2 km from the shoreline, flow heights were noted to be 3–4m foraminifera of large fossil e) % Abundance 0 above TP, and ~2 m above TP at a distance of 4 km from the shoreline (Goto et al., 2011). The hardest hit area was at a distance of 0–2.5 km 2 from the shoreline where houses and roads were severely damaged fi fl – 6 and rice elds ooded with saltwater. Co-seismic subsidence (17

D10 21 cm) approximately 10 km south of the Sendai airport was f) 10 reported (Ozawa et al., 2011) as well as earthquake-induced liquefac- tion in the adjacent rice fields (Goto et al., 2011). 2 The tsunami deposited a discontinuous unit that transitioned from 6 sand-dominated (0–2.8 km) to mud-influenced (2.8–4.5 km) sedi- D90 ments with increasing distance inland; Goto et al., 2011; Fig. 3). Ero- g) 10 sion was extreme where the tsunami breached an artificially 0 Standard deviation emplaced and reinforced dune system on the beach while the deposit was thickest in the coastal forest. Field descriptions of the tsunami 2 deposit document a landward thinning and texturally fining sand to mud deposit (Appendix 1). In some trenches, the sand deposit was 4 h) laminated with alternating sand and heavy mineral laminae (e.g. 1 trench 31; Fig. 2c). Small mollusk fragments were present in very b 0.5 minor amounts ( 1%) in intertidal and beach samples. They were also reported in the tsunami deposit but only rarely (0.2 and 0.8 km sphericity Increasing inland) and in minute amounts (b1%); Chagué-Goff et al., 2012–this i) 0 30 issue). Richmond et al. (2012–this volume) observed articulated bi-

20 valve shells west of Teizan-bori canal and concluded that they were probably derived from the canal, as none were found on the seaward 10 Deposit * erosional site side. 0 Thickness (cm) * 0.0 4.1 Differential volume % j) 4.1. Surface sediments 0 sand

7 silt Grain size results distinguished between intertidal and beach samples (sites 1, 2, 3), and dune (site 4) surface samples (Fig. 3). 14 clay – 012345 We found coastal sediments (sites 1 4) to have similar grain sizes Increasing distance inland (km) (D90 =0.0–1.4Φ), but varying degrees of sorting (StD=1.9–3.6Φ) and sphericity (0.3–0.9, with 1.0 being a perfect sphere). Intertidal Fig. 3. Surface samples (red) compared to changes in the Tōhoku-oki deposit (blue) sediment (average D90 =0.5Φ), with the lowest sphericity values with increasing distance inland. (a) Elevation along a transect from the intertidal fi fi (0.4), was most angular in composition, followed by beach (D90 = coastline of the Paci c ocean through the coastal forest, paved landscape and rice elds Φ fi Φ using the Tokyo Peil datum (TP; mean sea level in Tokyo Bay; see Fig. 1a, c). Measured 0.0 ; sphericity=0.8) and arti cial dune sediments (D90 =1.4 ; tsunami flow heights are indicated in gray. (b–d) Total concentration and relative sphericity=0.9), which were significantly more rounded. – abundances of foraminiferal taphonomic data. (e h) Average D10 (e), D90 (f), standard Surface sediment samples also showed distinctly different forami- deviation (sorting; g), and degree of angularity (sphericity; h) data for all Tōhoku-oki niferal characteristics related to increasing distance away from the deposit intervals in a single trench. Due to the dominant silt component of the deposit marine source. Intertidal samples (sites 1 and 2) had the lowest con- at trenches 82 and 86, sphericity could not be assessed. (i) Tsunami deposit thickness. 3 3 Trench 19 was recovered from an area where the tsunami deposited little to no sedi- centration of recent individuals per cm (9 per cm ), the highest con- 3 ment but washed away existing playground equipment. (j) Average particle size distri- centration of fossils (124 per cm ) and the greatest abundance of bution (PSD) plot of all tsunami deposit intervals in a single trench. large fossil specimens (59 per cm3). The artificial dune (site 4) was characterized by the lowest concentration of fossils (19 individuals per cm3), but the highest concentration of recent foraminifera (77 J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89 83 individuals per cm3), which were comparatively small in size (only 4.2. Lateral changes within the Tōhoku-oki tsunami deposit 32% of fossils were >250 μm). Beach sediment (site 3) marked a transition zone and had intermediate concentrations (fossil concentra- The tsunami deposit showed trends with increasing distance in- tion=90 individuals per cm3; recent concentration=2 individuals land (Fig. 3b–j). In general, sediments become finer grained (average 3 per cm ; % large fossils=59%). Intertidal and beach samples were D10 =1.8Φ at 0.3 km; 7.3Φ at 3.0 km; 8.1Φ at 4.5 km), less sorted predominantly composed of low abundances of miliolids (~4/cm3; (average StD=2.3Φ at 0.3 km; 2.2Φ at 3.0 km; 2.9Φ at 4.5 km) and e.g. Quinqueloculina spp., Triloculina spp; Plate 1 specimens 7–10) and more angular (0.7 at 0.3 km; 0.3 at 3.0 km; 0.4 at 3.7 km) with in- Ammonia parkinsoniana (2/cm3), while the dune contained only mar- creasing distance inland. These results are slightly coarser than ginally higher abundances (A. parkinsoniana=20/cm3, Quinequloculina those reported by Szczuciński et al. (2012–this volume) and can be spp.=6/cm3, miliolids=51/cm3). attributed to differences in sampling resolution. Medium to coarse

Fossil specimens 1 2 34

Recent specimens

56 Highly corroded tests with rounded edges

7 8 910

Plate 1. All scale bars are equal to 100 μm. (1–2) Light microscope images of sediment in-ﬁlled fossil specimens. (3–4) SEM images of fossil specimens indicating highly corroded and abraded tests. (5) Recent A. parkinsoniana ventral view. (6) Recent A. parkinsoniana dorsal view. (7–10) Taphonomically altered (corroded, abraded, edge rounded) recent miliolids. 84 J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89

sand was generally the dominant (e.g. D90) particle size throughout decrease in abundance at trench 48 (24%; 12% Amphistegina spp.; the entire deposit and only ranged from 0 to 2Φ; however, the fine 28% miliolids) and are non-existent by trench 86 (Fig. 3d). component (e.g. D10) of the deposit varied considerably with D10 values varying by 6.2Φ (Fig. 3e, f, j; Appendix 1). The deposit thick- 4.3. Vertical changes within the Tōhoku-oki tsunami deposit ness also thinned from 25 cm at a distance of 1 km from the shoreline to b5 mm 4.5 km inland (Fig. 4). The six trench sections (5, 12, 31, 36, 48, 86) are characterized by three The tsunami deposit contained a combination of recent (e.g. calcar- distinct units (Figs. 4 and 5). Trenches had basal rice field soil (except at eous, late Holocene) and fossil (sediment in-filled, calcified, Miocene to trench 5 where there was a basal forest soil) with a pronounced Pleistocene) foraminifera. Recent foraminifera were taphonomically yellow-brown color, consisting of poorly sorted to very poorly sorted me- altered showing signs of significant fragmentation, edge rounding dium sand. This was sharply overlain with a medium‐grained sand depos- (abrasion) and dissolution. Fractured edges also showed evidence of it that transitioned into a muddy-sand unit consisting of clay to fine sand edge rounding indicating fragmentation occurred before tsunami depo- sized particles. The fine component of the muddy-sand layer increased sition. Taphonomic alteration prevented proper species identification with distance inland and ultimately became the dominant sedimentary except A. parkinsoniana, which was present in most samples, although unit ranging from very fine sand (D10 =3.8Φ) 1.6 km inland to clay 3 in very low abundances (b20 individuals per 1 cm ; Plate 1). Miliolids (D10 =8.1Φ) at ~4 km. The muddy-sand layer was evident at sites greater were also found in low abundances. Analyses of the trench sections ver- than 1.2 km from the coastline, but was often discontinuous. The sand sus distance inland showed analogous relations to the surface samples and muddy-sand units (where present) together comprise the tsunami regarding the abundances of fossil and recent foraminifera. Recent indi- deposit. The tsunami deposit generally fined upward and became less viduals, although low in abundance, peaked at trenches 5 (~25 individ- sorted, and in some cases (e.g. trench 31) contained finer dark laminae uals per cm3) and 12 (~29 individuals per cm3) and decreased by ~50% (Fig. 5). Particle sphericity did not show any consistent vertical trends by trench 48 (~12 individuals per cm3). At the landward extent of our within the tsunami sands or between the tsunami sands and the sediment transect (~4.5 km, trench 86) no recent foraminifera were muddy-sand. found (Fig. 3c). Foraminifera (fossil and recent) are absent within the soil, except Fossil foraminifera were more robust, calcified, darker in color, at trenches 5 and 31 where very low abundances (11 recent and 65 highly abraded and more abundant than recent specimens within fossil individuals per cm3 at trench 5; 3 recent and 53 fossil individ- the tsunami deposit in all trenches (Fig. 5). Abundances of fossil fora- uals per cm3 at trench 31) are found near the contact with the over- minifera within the tsunami deposit peak at trenches 9 and 12 (~102 lying sand suggesting some bioturbation (Fig. 5). Foraminifera are individuals per cm3, ~104 individuals per cm3 respectively) immedi- present in the tsunami sand (19 recent and 82 fossil individuals per ately inland of the coastal forest (0.4–0.5 km inland), decrease to less cm3) and muddy-sand (8 recent and 10 fossil individuals per cm3), than 3 individuals per cm3 by trench 86 (Fig. 3b) and contain relative- with little or no variations in abundance with depth, except at trench- ly high abundances of Amphistegina spp. (e.g. individual 1 in Plate 1), es containing a muddy sand-layer where abundances of fossil speci- Quinqueloculina spp., and unrecognizable miliolids (Appendices 2, 3). mens are significantly higher in the sand than in the overlying Similarly, large fossil individuals (>250 μm) dominate the area be- muddy-sand. However, the proportion of large recent and fossil fora- tween trench 12 (67%; 22% Amphistegina spp.; 43% miliolids) and minifera was highest at the bottom of the tsunami deposit and trench 31 (64%; 13% Amphistegina spp.; 40% miliolids), rapidly showed a slight upward fining sequence in most trenches. For

a) c) 2011 A.D. 869 40 0 Towada-a tephra

14 0 Φ 14 0 Φ 10 50 Trench depth (cm) 0.0 5.0 14 0 Φ D.V. % 0.0 6.5 D.V.% = Differential volume (%) D.V. % Sample interval Soil 0.0 6.2 D.V. % d) 20 N 14 0 Φ b) 36 31 Muddy-sand 48 12 0.0 5.3 14 0 Φ 500 m D.V. % Φ Sendai airport 5 Grain size ( ) 14 0 Φ Basal soil Agricultural land Beach/artificial dune Low density housing 0.0 8.2 0.0 6.7 Coastal forest D.V. % D.V. % Canal Trench location 5 1231 36 48 No. Trench 0 123 Increasing distance inland (km)

Fig. 4. (a) Particle size distribution (PSD) plots for Tōhoku-oki trench sections along the transect. Facies designations are based on ﬁeld observations and black dots represent sampling intervals. (b) Generalized stratigraphic section of the Tōhoku-oki tsunami deposit from 1.5 to 2.8 km inland. (c) PSD plot for the Jōgan trench section including the overlying soil and Towada-a tephra deposited by volcanic activity to the north of Sendai in A.D. 870–934 (Fig. 1a). (d) Location of trench sites. For elevation (meters above TP) and distance from the coastline see Fig. 3a. J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89 85

muddy-sand shows slightly elevated TOC values (0.3%) and much higher δ13C(−15.1‰). Grain size 4.4. A.D. 869 Jōgan tsunami deposit

Trench 5 13 10 90 δ C (‰) 0 Total individualsTotal perrecentTotal 1 cm foraminifera fossil% Largeforaminifera per recent% 1Large cm perforaminifera fossil 1D cm foraminiferaD Std (sorting)Sphericity We compared the Tōhoku-oki deposit with an older event of similar -32 -20 magnitude, the A.D. 869 Jōgan tsunami (Fig. 1). The Jōgan trench consisted of ﬁve stratigraphic units (basal soil, Jōgan tsunami deposit, 10 soil, Towada-a tephra, and overlying soil; Fig. 4c), of which the basal soil, tsunami deposit and soil were sampled and analyzed. The basal

Depth (cm) soil is composed of a very poorly sorted (StD=2.8Φ)sandysoil 20 δ13C %C (D10=5.2Φ, D90=1.1Φ). The overlying Jōgan tsunami deposit is a 10‐ 50 50 3 6 150 300 200 100 100 200 100 100 4.5 8.5 1.0 2.5 1.7 3.5 0.5 1.0 ‐ Φ Φ TOC cm thick very poorly sorted (StD=3.0 ), medium sand (D90=1.0 ). (wt/wt) Trench 12 The tsunami deposit is overlain by soil which is itself overlain by the 0 Towada-a tephra (D10=6.4Φ;StD=2.7Φ). The contacts between these four units were gradational. The Jōgan tsunami deposit showed similar sedimentological characteristics as the Tōhoku-oki: slight ﬁning 10 in grain size, better sorting, and greater particle sphericity (Fig. 4)from the bottom of the deposit to the top. The Jōgan tsunami deposit showed ﬂ 20 a pronounced in ux of highly spherical sediment (sphericity=0.7) compared to the overlying Towada-a (0.3) and underlying soil (0.3;

50 50 Figs. 3hand5). In the modern environment, highly spherical sediments 150 300 200100 100200 100 100 4.5 8.5 1.0 2.5 1.7 3.5 0.5 1.0 Trench 31 13 δ C (‰) seem to be originating from the beach (0.8) and dunes (0.9). 0 -28 -16 Foraminifera were present (~190 recent and ~162 fossil individuals per cm3) in the tsunami deposit indicating a probable intertidal origin. Low abundances of foraminifera were found in the upper sam- 10 ples of the basal soil (fossil=42; recent=65 individuals per cm3) 50 50 13 150 300 200100 100200 100 100 4.5 8.5 1.0 2.5 1.7 3.5 0.5 1.0 δ C %C and the Towada-a tephra (fossil=3; recent=16 individuals per 3 6 3 Trench 48 TOC cm ) suggesting bioturbation. This is similar to basal soils underlying (wt/wt) 0 the Tōhoku-oki deposit (trenches 5, 31 and 48), where no recent or fossil foraminifera were found below the bioturbated contact (Fig. 5). The recent foraminiferal assemblage of the Jōgan tsunami de- 10 50 50 150 300 100 200 100 200 100 100 4.5 8.5 1.0 2.5 1.7 3.5 0.5 1.0 posit consisted of A. parkinsoniana, various taphonomically altered ō Jogan trench miliolids and unaltered planktics. This is in contrast to the T hoku-oki Towada-a tephra ō 40 deposit where no planktics were found. Compared to the T hoku-oki deposit, A. parkinsoniana individuals were more altered, showing signs of increased abrasion and dissolution. Large test sizes dominate 50 the tsunami deposit and do not appear to show evidence of grading.

5. Discussion 60 50 50 150 300 200100 200100 100 100 4.5 8.5 1.0 2.5 1.7 3.5 0.5 1.0 No. of Abundance Phi (Φ) individuals (%) 5.1. Stratigraphy and grain size analyses of the 2011 Tōhoku-oki tsunami

Fig. 5. Grain size and foraminiferal taphonomic data for trenches 5, 12, 31, 48 and Tsunami deposits of Hokkaido Japan (Sawai, 2002), New Zealand Jōgan. δ13C, TOC data for trenches 5 and 31 are shown. Black dots indicate sampling (Goff et al., 2001), Papua New Guinea (Morton et al., 2007), Cascadia intervals. (Hawkes et al., 2011) and elsewhere have been described on the basis of their lateral, sheet-like geometry, with the deposit thickness taper- ing inland (Goff et al., 2001; Morton et al., 2007). At Sendai, the tsu- example, recent and fossil foraminifera at the base of the tsunami de- nami deposit was variable in thickness, extended over a 4.5 km posit at trench 12 were 63% and 71% large (>250 μm) respectively transect, tapered inland from 25 cm to less than 5 mm and contained and decreased to 45% and 56% at the top of the unit. Test size grading fine dark laminae interbedded with sand at some trenches (trench was most pronounced in trenches containing capping muddy-sand 31), as also reported by Szczuciński et al. (2012–this volume) and layers (e.g. trench 31: 70% large sized fossils at the bottom of the tsu- Richmond et al. (2012–this volume). The Tōhoku-oki sand deposit is nami unit, 50% at the top of the sand and 30% in the muddy-sand similar in grain size (average D10 =4.0Φ, average D90 =0.7Φ) and de- layer; 70% large sized recent foraminifera at the bottom, 49% at the gree of sorting (StD=2.2Φ) to modern intertidal (average D10 = top of the tsunami sand and 40% in the muddy-sand). 1.9Φ, average D90 =0.5Φ; StD=2.2Φ), beach (D10 =1.5Φ; D90 = 13 Total organic carbon and δ C of two trench sections (5, 31) distin- 0.0Φ; StD=1.9Φ) and dune (D10 =3.7Φ; D90 =1.4Φ; StD=3.6Φ) guished between the tsunami deposit and the underlying soil (Fig. 5). surface samples (Fig. 3e–g) supporting the suggestion of Goto et al. In trench 5, δ13C ranged from −27.0 to −24.8‰ in the tsunami sand (2011) and Szczuciński et al. (2012–this volume) who ascribe a near- and −29.5 to −30.8‰ in the soil. TOC values were notably low in the shore to dune origin for the tsunami sand. tsunami deposit (~0.1%) compared with the underlying soil (0.5%– The finer underlying rice field soil (average D10 =4.6Φ; average 5.9%). Similar to grain size results, three distinct units are distin- D90 =0.5Φ) sharply transitioned to a coarser sand deposit (average 13 guished in the TOC and δ C profile of trench 31. From 17.5 to D10 =4.0Φ; average D90 =0.7Φ) that, in some cases, fined upward 13 11.5 cm the soil has δ C values of −27.8‰ to −27.1‰ and TOC to a muddy-sand layer (average D10 =7.2Φ; average D90 =1.3Φ). values ranging from 1.5% to 5.7%. The TOC decreases to 0.1% and the PSD data augmented statistical grain size values by indicating a dis- δ13C values increase to −26.3‰ in the tsunami sand. The tsunami tinct fine tail at the top of trenches 31, 36 and 48 (Fig. 4). The fining 86 J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89 upward sequence within the Tōhoku-oki deposit is in agreement with in surface intertidal sediments. Miocene to Pleistocene marine sedi- several other studies (e.g. Hawkes et al., 2007; Morton et al., 2007; ments (unconsolidated) and sedimentary rock surrounding the Sendai Goodman-Tchernov et al., 2009) and represents entrainment of sedi- plain are the probable source of the fossil foraminifera (Fig. 1b). The cal- ment followed by rapid deposition. The muddy-sand layer that caps cified tests of fossil individuals indicate diagenesis and favor a lithified the sand deposit at several trenches located at least 1.2 km from the rock origin over unconsolidated sedimentary units, of which there are shoreline up to 2.8 km and then becomes dominant represents fur- two main sources: Jurassic sedimentary rock outcropping on the Oshika ther waning energy and is likely derived from antecedent rice field Peninsula and near Sōma; and Early Cretaceous marine sedimentary soil, canal mud or deeper offshore mud. It is possible that the tsunami rock off Oshika. Because of the dominating northward longshore drift scoured deeper offshore (below storm wave base) entraining finer in Sendai Bay (Hattori, 1967), most fossil foraminifera are more likely grain sizes that were deposited in areas of sustained flooding. This in- to have originated from the south (area near Sōma). terpretation is further supported by TOC, δ13C and foraminiferal re- Both fossil and recent foraminifera abundances declined with in- sults (see below) which suggest an intertidal to beach origin for the creasing distance inland. The abundance of large sized individuals muddy-sand layer. Detailed sedimentary descriptions of the deposit (e.g. >250 μm) ranged from 60% to 70% between the coastline and are reported by Szczuciński et al. (2012–this volume) and 1.5 km inland where they began to markedly decrease to 30%–35% Richmond et al. (2012–this volume). at 2.5 km inland and finally to a negligible amount (e.g. b1%) at 4.5 km inland. The abrupt decrease in abundance of large sized indi- 5.2. Recent and fossil foraminifera as a tsunami indicator viduals at 1.2 km inland coincides with the beginning of the muddy-sand (e.g. trench 31 at 1.6 km and trench 48 at 2.4 km) and The presence of abundant foraminifera is a characteristic of tsuna- likely represents waning energy and sustained pooling of marine mi deposits (e.g. Hawkes et al., 2007; Kortekaas and Dawson, 2007; water. Mamo et al., 2009; Pilarczyk and Reinhardt, 2012a, 2012b), but at Sendai their abundance within the sand deposit was low. Hawkes et 5.3. TOC and δ13C trends al. (2007) found up to 1400 individuals per cm3 in the 2004 Indian Ocean tsunami deposit in Malaysia and Thailand, whereas we only Stable carbon isotopes (δ13C) and total organic carbon (TOC) have found up to 48 recent individuals per cm3. The lack of recent forami- been used extensively to infer the provenance of organic matter hosted nifera within intertidal sediments is also anomalous when compared in terrestrial, coastal wetland and marine sediments (e.g. Tyson, 1995; to Matoba (1976) who documented diverse foraminifera in the Lamb et al., 2007; Kemp et al., 2010; Vane et al., 2010). Although δ13C deeper areas of Sendai Bay (depths of 19–1988 m) and other studies and TOC have the potential to distinguish tsunami sediment from un- from other Japanese coastlines (e.g. Toba, Mie Prefecture, Hokkaido) derlying soils primarily because imported marine sands should have that report abundant and diverse assemblages (Okahashi et al., low TOC content and higher δ13C values than the local terrestrial soils, 2002; Nanayama and Shigeno, 2004, 2006). Szczuciński et al. they have seldom been applied. (2012–this volume) also report a surprising paucity of nannoliths In trench 5, the basal forest soil has low δ13Cvalues(−30.8‰)and (biogenic carbonate) and marine diatoms in intertidal areas as well TOC of up to 6%. Pine roots that are found in the soil have similar δ13C as in the Tōhoku-oki tsunami deposit. Furthermore, the foraminifera (−30.4‰), which are consistent with values reported from other forest within the tsunami deposit were highly taphonomically altered. Taxo- soils (e.g. Goni and Eglinton, 1996; Goni and Thomas, 2000; Vane et al., nomic identification was impossible except for A. parkinsoniana. A. 2003) and from the same area by Chagué-Goff et al. (2012–this issue). parkinsoniana has previously been documented as inhabiting brackish This suggests a woody terrestrial plant material source for the forest littoral (5–10 m deep) to sub-littoral (b300 m deep) areas in Hokkaido, soil. In contrast the overlying tsunami sand has δ13C values ranging Sanriku and Boso Peninsula, Japan (Takata et al., 2006; Uchida et al., from −27.0‰ to −24.8‰ and very low TOC of about 0.1% (Fig. 5). 2010) and elsewhere (e.g. Debenay et al., 1998), but not in Sendai Bay The δ13C of marine and open coastal sediments typically range between (Matoba, 1976). Low abundances of A. parkinsoniana are found −18‰ to −23‰ and estuarine sediments range from −26‰ to −23‰ unaltered in post-tsunami intertidal sediment and possible sources in- (Hedges and Mann, 1979; Mishima et al., 1999; Wilson et al., 2005). clude the brackish zone at the confluence of the canals and the sea, or Furthermore, surface sediments from the marine influenced section of allochthonous sediment that may have been brought in during or post Osaka Bay, Japan are characterized by δ13Cof−20‰ to −21‰ canal construction. (Mishima et al., 1999). In two trench sections containing the The taphonomic character of recent foraminifera have been used to Tōhoku-oki deposit we found the δ13C values of the sand unit to be document overwash deposits because it can provide additional infor- slightly more depleted in 13C (negative) than that expected for sedi- mation concerning energy regimes and transport history (e.g. Hawkes ment hosting purely marine-derived organic matter, but remain 4.0‰ et al., 2007; Kortekaas and Dawson., 2007; Satyanarayana et al., 2007; higher than the underlying soil. A similar contrast in TOC values be- Uchida et al., 2010; Pilarczyk et al., 2011; Pilarczyk and Reinhardt, tween the underlying soil and overlying tsunami deposit was more pro- 2012b). Within the Tōhoku-oki tsunami deposit, recent foraminifera nounced at trench 31 than trench 5. However the greatest δ13Cchange showed evidence of subaerial exposure through a high degree of is associated with the transition to the capping muddy-sand layer abrasion (edge rounding), corrosion and fragmentation (Berkeley (δ13C=26.3‰ to −15.1‰). The minor change in δ13Cbetweenthe et al., 2009; Pilarczyk et al., 2011). Contrary to other taphonomic basal soil and tsunami deposit at trench 31 probably reflects its studies of tsunami sediments (e.g. Kortekaas and Dawson, 2007), greater distance inland and thus a larger contribution of terrestrial fragmentation at Sendai was not a function of the tsunami. Rather, organic matter sources. The positive values of the muddy-sand may edge rounding of fractured surfaces indicates repeated subaerial ex- result from sediment containing organic matter from plants utilising posure and significant residence time in the coastal zone (beach and the C4 photosynthetic pathway (range −17 ‰ to −9‰; Deines, artificial dune). Abundances of recent foraminifera peak in modern 1980). Alternatively, the organic matter is sourced from either ma- dune samples and probably do not represent modern conditions at rine algae (−16 ‰to −24‰), marine plankton (−13 to −31‰), Sendai since the artificial dune sediment was transported from an marine particulate organic carbon (−18‰ to −24‰), marine bacteria unknown source. (−12‰ to −26‰), sea grasses (marine C4 plant; −14‰ to −19‰), Fossil individuals are found in all trenches, including the landward or possibly cyanobacteria (Deines, 1980; Tyson, 1995). The muddy- limit of the transect (4.5 km), and show a marked decrease in concen- sand unit shows evidence of organic matter of marine origin because tration and size with increasing distance inland. Fossil specimens C4 plants do not occur in isolation from C3 plants and would produce proved to be marine indicators because they were found predominantly amixedδ13C signature of about −21‰. The low TOC content of 0.3% J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89 87 suggests against input from plant organic matter which typically Acknowledgements increases overall a coastal sediments TOC content above 0.5%. The contribution of C4 agricultural plants (maize, sugar cane millet Members of the UNESCO-IOC (Intergovernmental Oceanographic sorghum) which also give overlapping δ13C can be largely discounted Commission) team to Japan provided samples and guidance. Among since these crops also tend to yield sediments with high δ13C and TOC them we especially thank Kazuhisa Goto and Bruce Jaffe for their in- values (1%–5% TOC) due to their cellulose and lignin rich structures. volvement in executing this field survey. We also thank Daisuke Sugawara, Yuichi Nishimura and Shigehiro Fujino for their assis- tance while on site in Japan, Kazuhisa Goto for providing survey 5.4. Comparison between the Tōhoku-oki and Jōgan tsunami deposits data and Doug Jerolmack for the use of his camsizer. C. Vane pub- lishes with the permission of the Executive Director, British Geolog- Historical records mention several tsunamis that have impacted ical Survey (NERC). The work was supported in part by UNESCO-IOC, northeast Japan (Minoura and Nakaya, 1991; Minoura et al., 2001). The U.S. Geological Survey and a National Science Foundation award However, the A.D. 869 Jōgan tsunami is reported to be similar to the (EAR-0842728). This paper is a contribution to IGCP 588. Tōhoku-oki event in terms of the extent of inundation (Goto et al., 2011). Both tsunamis deposited a laterally extensive, landward thinning sand deposit that extended to distances greater than 2 km References from the coast (Goto et al., 2011; Sugawara et al., 2011b). Association of Japanese Geographers, 2011. Map of tsunami inundation reported by Tsunami Grain size distributions were effective in discriminating the sand Damage Team. http://danso.env.nagoya-u.ac.jp/20110311/map/574017Sendaikukou. units (medium sand) deposited by both the Tōhoku-oki and Jōgan jpg. fi ō Berkeley, A., Perry, C.T., Smithers, S.G., 2009. Taphonomic signatures and patterns of tsunamis from the ner soils. At the studied trench, the J gan sand test degradation on tropical, intertidal benthic foraminifera. Marine Micropaleon- deposit could be distinguished from the underlying basal soil and tology 73, 148–163. overlying Towada-a tephra even though contacts were gradational. Bernard, E.N., Robinson, A.R., 2009. Tsunamis. The Sea, vol. 15. Harvard University ō Press, Cambridge, MA. 450 pp. Both tsunami deposits consisted of very poorly sorted (T hoku-oki: Blott, S.J., Pye, K., 2001. Gradistat: a grain size distribution and statistics package for the 2.0Φ;Jōgan: 3.0Φ) medium sand (Tōhoku-oki: D90 =0.7Φ;Jōgan: analysis of unconsolidated sediments. Earth Surface Processes and Landforms 26, – D90 =1.0Φ) that fined upward (Fig. 4). Notable differences between 1237 1248. the deposits include the absence of a muddy-sand layer and lami- Bondevik, S., Lovholt, F., Harbitz, C., Mangerud, J., Dawson, A., Svendsen, J.I., 2005. The Storegga Slide tsunami—comparing field observations with numerical observa- nae in the Jōgan deposit. This may be evidence of depositional tions. Marine and Petroleum Geology 22, 195–208. style (e.g. magnitude of earthquake, sediments available for trans- Chagué-Goff, C., 2010. Chemical signatures of palaeotsunamis: a forgotten proxy? Ma- – port, flow depth, etc.), or post-depositional change whereby bioturba- rine Geology 271, 67 71. ń Chagué-Goff, C., Dawson, S., Goff, J., Zachariasen, J., Berryman, K., Garnett, D., Waldron, tion obscures internal structures (Szczuci ski, 2012). The absence of a H., Mildenhall, D., 2002. A tsunami (ca. 6300 years BP) and other Holocene envi- muddy-sand layer could also be related to the location of the Jōgan ronmental changes, northern Hawke's Bay, New Zealand. Sedimentary Geology trench, that was sampled 1.5 km from the present coastline, which is 150, 89–102. Chagué-Goff, C., Schneider, J., Goff, J.R., Dominey-Howes, D., Strotz, L., 2011. Expanding equivalent to ~0.5 km from the paleo-coastline. No muddy-sand units the proxy toolkit to help identify past events—lessons from the 2004 Indian Ocean were reported from the Tōhoku-oki tsunami within 1.2–1.5 km from tsunami and the 2009 South Pacific tsunami. Earth-Science Reviews 107, 107–122. the shoreline. Chagué-Goff, C., Andrew, A., Szczuciński, W., Goff, J., Nishimura, Y., 2012. Geochemical signatures up to the maximum inundation of the 2011 Tohoku-oki tsunami— The recent and fossil foraminifera also showed broad similarities implications for the 869 AD Jogan and other palaeotsunamis. Sedimentary Geology and noticeable differences. Foraminifera were more abundant within 282, 65–77 (this issue). the Tōhoku-oki and Jōgan tsunami deposits although bioturbation Chagué-Goff, C., Niedzielski, P., Wong, H.K.Y., Szczuciński, W., Sugawara, D., Goff, J., 2012. Environmental impact assessment of the 2011 Tohoku-oki tsunami on the may have resulted in their occurrence in the underlying soil, and in Sendai plain. Sedimentary Geology 282, 175–187 (this issue). the case of Jōgan, also the overlying units. In the Jōgan tsunami deposit Cisternas, M., Atwater, B.F., Torrejon, F., Sawai, Y., Machuca, G., Lagos, M., Eipert, A., concentrations of fossil and recent individuals were similar (recent= Youlton, C., Salgado, I., Kamataki, T., Shishikura, M., Rajendran, C.P., Malik, J.K., 190 individuals per cm3; fossil=162 individuals per cm3), which is in Rizal, Y., Husni, M., 2005. Predecessors of the giant 1960 Chile earthquake. Nature 437, 404–407. contrast to the Tōhoku-oki deposit where fossil foraminifera are more Debenay, J., Beneteau, E., Zhang, J., Stouff, V., Geslin, E., Redois, F., Fernandez-Gonzalez, abundant. The recent assemblage of foraminifera in the Jōgan sequence M., 1998. Ammonia beccarii and Ammonia tepida (Foraminifera): morphofunctional – was dominated by A. parkinsoniana and miliolids, but there was a no- arguments for their distinction. Marine Micropaleontology 34, 235 244. Deines, P., 1980. The isotopic composition of reduced organic carbon. In: Fritz, P., ticeable presence of planktic individuals. Planktic foraminifera have Fontes, J.C. (Eds.), Handbook of Environmental Isotope Geochemistry. Elsevier, also been used to document tsunami sediments from the Indian Amsterdam, pp. 329–426. Ocean (Hawkes et al., 2007), the Aegean Sea (Minoura et al., 2012) Donato, S.V., Reinhardt, E.G., Boyce, J.I., Pilarczyk, J.E., Jupp, B.P., 2009. Particle-size distribution of inferred tsunami deposits in Sur Lagoon, Sultanate of Oman. Marine and the Arabian Sea (Pilarczyk and Reinhardt, 2012b). Geology 257, 54–64. Geological Survey of Japan, AIST (ed.), 2009. Seamless digital geological map of Japan 1:200,000. December 15, 2009 version. Research Information Database DB084, Geological Survey of Japan, National Institute of Advanced Industrial Science and 6. Conclusions Technology. Goff, J., Chagué-Goff, C., 1999. A late Holocene record of environmental changes from The 2011 Tōhoku-oki tsunami shows marked trends of inland fin- coastal wetlands: Abel Tasman National Park, New Zealand. Quaternary International 56, 39–51. ing and thinning and can be discriminated from the underlying soil by Goff, J., Chagué-Goff, C., Nichol, S., 2001. Palaeotsunami deposits: a New Zealand per- an abrupt increase in grain size (medium sand) and the notable ap- spective. Sedimentary Geology 143, 1–6. pearance of recent and fossil foraminifera. TOC and δ13C analysis of Goff, J., Chagué-Goff, C., Dominey-Howes, D., McAdoo, B., Cronin, S., Bonté-Grapetin, the Tōhoku-oki deposit also revealed a contrast between the soil M., Nichol, S., Horrocks, M., Cisternas, M., Lamarche, G., Pelletier, B., Jaffe, B., Dudley, W., 2011. Palaeotsunamis in the Pacific. Earth-Science Reviews 107, and the overlying tsunami unit at one trench and corroborates grain 141–146. size and foraminiferal results that indicate an intertidal source for Goni, M.A., Eglinton, T.I., 1996. Stable carbon isotopic analyses of lignin-derived CuO the sand. Problems of identification of recent foraminifera along the oxidation products by isotope ratio monitoring gas chromatography mass spec- trometry (irm–GC–MS). Organic Geochemistry 24 (6–7), 601–615. Sendai transect limit their utility; however, the added use of taphon- Goni, M.A., Thomas, K.A., 2000. Sources and transformations of organic matter in sur- omy (test condition, fossil specimens) helped to constrain sediment face soils and sediments from a tidal estuary (north inlet, South Carolina, USA). provenance and hydrodynamic regime. Estuaries 23 (4), 548–564. Goodman-Tchernov, B.N., Dey, H.W., Reinhardt, E.G., McCoy, F., Mart, Y., 2009. Tsunami Supplementary data to this article can be found online at http:// waves generated by the Santorini eruption reached eastern Mediterranean shores. dx.doi.org/10.1016/j.sedgeo.2012.08.011. Geology 37, 943–946. 88 J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89

Goto, K., Chagué-Goff, C., Fujino, S., Goff, J., Jaffe, B., Nishimura, Y., Richmond, B., Ozawa, S., Nishimura, T., Suito, H., Kobayashi, T., Tobita, M., Imakiire, T., 2011. Sugawara, D., Szczucinski, W., Tappin, D.R., Witter, R.C., Yulianto, E., 2011. New in- Coseismic and postseismic slip of the 2011 magnitude 9 Tohoku-oki earthquake. sights of tsunami hazard from the 2011 Tōhoku-oki event. Marine Geology 290, Nature 475, 373–376. 46–50. Patterson, R.T., Fishbein, E., 1989. Re-examination of the statistical methods used to de- Hattori, M., 1967. Recent sediments of Sendai Bay, Miyagi Prefecture, Japan. Scientific termine the number of point counts needed for micropaleontological quantitative Reports of Tohoku University, 2nd Series (Geology), 39, pp. 1–61. research. Journal of Paleontology 63 (2), 245–248. Hawkes,A.D.,Bird,M.,Cowie,S.,Grundy-Warr,C.,Horton,B.P.,TanShauHwai,A., Pilarczyk, J.E., Reinhardt, E.G., 2012a. Homotrema rubrum (Lamarck) taphonomy as an Law,L.,Macgregor,C.,Nott,J.,EongOng,J.,Rigg,J.,Robinson,R.,Tan-Mullins, overwash indicator in marine ponds on Anegada, British Virgin Islands. Natural M., Tiong, T., Yasin, Z., Wan Aik, L., 2007. Sediments deposited by the 2004 Indian Hazards 63, 85–100. Ocean tsunami along the Malaysia–Thailand Peninsula. Marine Geology 242, Pilarczyk, J.E., Reinhardt, E.G., 2012b. Testing foraminiferal taphonomy as a tsunami in- 169–190. dicator in a shallow arid system lagoon: Sur, Sultanate of Oman. Marine Geology Hawkes, A.D., Horton, B.P., Nelson, A.R., Vane, C.H., Sawai, Y., 2011. Coastal subsidence 295–298, 128–136. in Oregon, USA, during the giant Cascadia earthquake of AD 1700. Quaternary Pilarczyk, J.E., Reinhardt, E.G., Boyce, J.I., Schwarcz, H.P., Donato, S.V., 2011. Assessing Science Reviews 30, 364–376. surficial foraminiferal distributions as an overwash indicator in Sur Lagoon, Sultan- Hayward, B.W., Holzmann, M., Grenfell, H.R., Pawlowski, J., Triggs, C.M., 2004. Morpho- ate of Oman. Marine Micropaleontology 80, 62–73. logical distinction of molecular types in Ammonia—towards a taxonomic revision Pinegina, T.K., Bourgeois, J., Bazanova, L.I., Melekestsev, I.V., Braitseva, O.A., 2003. A of the world's most commonly misidentified foraminifera. Marine Micropaleontol- millennial-scale record of Holocene tsunamis on the Kronotskiy Bay coast, Kamchatka, ogy 50, 237–271. Russia. Quaternary Research 59, 36–47. Hedges, J.I., Mann, D.C., 1979. The lignin geochemistry of marine sediments from the Richmond, B., Szczuciński, W., Chagué-Goff, C., Goto, K., Sugawara, D., Witter, R., Tappin, southern Washington coast. Geochimica et Cosmochimica Acta 43, 1809–1818. D.R., Jaffe, B., Fujino, S., Nishimura, Y., Goff, J., 2012. Erosion, deposition and landscape Hemphill-Haley, E., 1995. Diatom evidence for earthquake-induced subsidence and change on the Sendai coastal plain, Japan resulting from the March 11, 2011 Tohoku- tsunami 300 years ago in southern coastal Washington. Geological Society of oki tsunami. Sedimentary Geology 282, 27–39 (this volume). America Bulletin 107 (3), 367–378. Saito, Y., 1991. In: MacDonald, D.I.M. (Ed.), Sequence stratigraphy on the shelf and upper Horton, B.P., Edwards, R.J., 2006. Quantifying Holocene sea level change using intertidal slope in response to the latest Pleistocene–Holocene sea-level changes off Sendai, foraminifera: lessons from the British Isles. Cushman Foundation for Foraminiferal northeast Japan, Sedimentation, Tectonics and Eustacy, Sequence Stratigraphy: Inter- Research, Special Publication, 40. 97 pp. national Assocaition of Sedimentologists Special Publication, 12, pp. 133–150. Ito, Y., Tsuji, T., Osada, Y., Kido, M., Inazu, D., Hayashi, Y., Tsushima, H., Hino, R., Sambridge, M., Braun, J., McQueen, H., 1995. Geophysical parametrization and interpo- Fujimoto, H., 2011. Frontal wedge deformation near the source region of the lation of irregular data using natural neighbors. Geophysical Journal International 2011 Tohoku-oki earthquake. Geophysical Research Letters 38 http://dx.doi.org/ 122, 837–857. 10.1029/2011GL048355. Satake, K., Namegaya, Y., Yamaki, S., 2008. Numerical simulation of the A. D. 869 Jogan Jankaew, K., Atwater, B.F., Sawai, Y., Choowong, M., Charoentitirat, T., Martin, M.E., tsunami in Ishinomaki and Sendai Plains. Annual Report of Active Fault and Prendergast, A., 2008. Medieval forewarning of the 2004 Indian Ocean tsunami in Paleoearthquake Research 8, 71–89. Thailand. Nature 455, 1228–1231. Sato, M., Ishikawa, T., Ujihara, N., Yoshida, S., Fujita, M., Mochizuki, M., Asada, A., 2011. Kelsey, H.M., Nelson, A.R., Hemphill-Haley, E., Witter, R.C., 2005. Tsunami history of an Displacement above the hypocenter of the 2011 Tohoku-oki earthquake. Science Oregon coastal lake reveals a 4600 yr record of great earthquakes on the Cascadia 332, 1395. subduction zone. GSA Bulletin 117 (7–8), 1009–1032. Satyanarayana, K., Reddy, A.N., Jaiprakash, B.C., Chidambaram, L., Srivastava, S., Kemp, A.C., Vane, C.H., Horton, B.P., Culver, S.J., 2010. Stable carbon isotopes as poten- Bharktya, D.K., 2007. A note on foraminifera, grain size and clay mineralogy of tsu- tial sea-level indicators in salt marshes, North Carolina, USA. The Holocene 20 (4), nami sediments from Karaikal–Nagore–Nagapattiuam Beaches, Southeast Coast of 623–636. India. Journal Geological Society of India 69, 70–74. Koper, K.D., Hutko, A.R., Lay, T., Ammon, C.J., Kanamori, H., 2011. Frequency-dependent Sawai, Y., 2002. Evidence for 17th-century tsunamis generated on the Kuril–Kamchatka rupture process of the 2011 Mw 9.0 Tohoku-oki Earthquake: comparison of short- subduction zone, Lake Tokotan, Hokkaido, Japan. Journal of Asian Earth Sciences 20, period P wave backprojection images and broadband seismic rupture models. 903–911. Earth Planets Space 63, 599–602. Sawai, Y., Fujii, Y., Fujiwara, O., Kamataki, T., Komatsubara, J., Okamura, Y., Satake, K., Kortekaas, S., Dawson, A.G., 2007. Distinguishing tsunami and storm deposits: an ex- Shishikura, M., 2008a. Marine incursions of the past 1500 years and evidence of ample from Martinhal, SW Portugal. Sedimentary Geology 200, 208–221. tsunamis at Suijin-numa, a coastal lake facing the Japan Trench. The Holocene Lamb, A.L., Vane, C.H., Wilson, G.P., Rees, J.G., Moss-Hayes, V.L., 2007. Assessing δ13C 18, 517–528. and C/N ratios from organic material in archived cores as Holocene sea level and Sawai, Y., Shishikura, M., Komatsubara, J., 2008b. A study of paleotsunami using had corer palaeoenvironmental indicators in the Humber Estuary, UK. Marine Geology 244, in Sendai plain (Sendai City, Natori City, Iwanuma City, Watari Town, Yamamoto 109–128. Town), Miyagi, Japan. Annual Report of Active Fault and Paleoearthquake Researches, Loeblich, A.R., Tappan, H., 1987. Foraminiferal Genera and Their Classification. Van 8, pp. 17–70 (in Japanese with English abstract). Nostrand Rienhold Co., New York. Sugawara, D., Imamura, F., 2010. Geological and historical traces of the 869 Jogan tsu- Machida, H., Arai, F., 2003. Atlas of Tephra in and Around Japan, revised edition. University nami field guide of half-day field trip in Sendai. The 3rd International Tsunami of Tokyo Press, Tokyo . 336 pp. Field Symposium, Sendai, Sanriku and Okinawa, 10–16 April 2010. Mamo, B., Strotz, L., Dominey-Howes, D., 2009. Tsunami sediments and their foraminif- Sugawara, D., Goto, K., Chagué-Goff, C., Fujino, S., Goff, J., Jaffe, B., Nishimura, Y., eral assemblages. Earth-Science Reviews 96, 263–278. Richmond, B., Szczucinki, W., Tappin, D.R., Witter, R., Yulianto, E., 2011a. Initial Matoba, Y., 1976. Recent foraminiferal assemblages off Sendai, northeast Japan. In: field survey report of the 2011 East Japan Tsunami in Sendai, Natori and Iwanuma Schafer, C.T., Pelletier, B.R. (Eds.), First International Symposium on Benthic Fora- cities. UNESCO-IOC International Tsuanmi Survey Team . 16 pp. minifera of Continental Margins, Part A Ecology and Biology: Maritime Sediments, Sugawara, D., Imamura, F., Matsumoto, H., Goto, K., Minoura, K., 2011b. Reconstruction Special Publication, No. 1, pp. 205–220. of the AD 869 Jogan earthquake induced tsunami by using the geological data. Minoura, K., Nakaya, S., 1991. Traces of tsunami preserved in inter-tidal lacustrine and Journal of Japanese Society of Natural Disaster Science 29, 501–516. marsh deposits: some examples from northeast Japan. Journal of Geology 99, 265–287. Sugawara, D., Imamura, F., Goto, K., Matsumoto, H., Minoura, K., 2012. The 2011 Minoura, K., Nakaya, S., Uchida, M., 1994. Tsunami deposits in a lacustrine sequence of Tohoku-oki earthquake tsunami: similarities and differences to the 869 Jogan tsu- the Sanriku coast, northeast Japan. Sedimentary Geology 89, 25–31. nami on the Sendai plain. Pure and Applied Geophysics http://dx.doi.org/10.1007/ Minoura, K., Imamura, F., Sugawara, D., Kono, Y., Iwashita, T., 2001. The 869 Jogan tsu- s00024-012-0460-1. nami deposit and recurrence interval of large-scale tsunami on the Pacific coast of Szczuciński, W., 2012. The post-depositional changes of the onshore 2004 tsunami de- northeast Japan. Journal of Natural Disaster Science 23, 83–88. posits on the Andaman Sea coast of Thailand. Natural Hazards 60, 115–133. Minoura, K., Imamura, F., Kuran, U., Nakamura, T., Papadopoulos, G.A., Takahashi, T., Szczuciński, W., Kokociński, M., Rzeszewski, M., Chagué-Goff, C., Cacho, M., Goto, K., Yalciner, A.C., 2012. Discovery of Minoan tsunami deposits. Geology 28 (1), 59–62. Sugawara, D., 2012. Sediment sources and sedimentation processes of 2011 Mishima, Y., Hoshika, A., Tanimoto, T., 1999. Depositional rates of terrestrial and ma- Tohoku-oki tsunami deposits on the Sendai Plain, Japan — Insights from diatoms, rine organic carbon in the Osaka Bay, Seto Inland Sea, Japan determined using car- nannoliths and grain size distribution. Sedimentary Geology 282, 40–56 (this bon and nitrogen stable isotope ratios in the sediment. Journal of Oceanography volume). 55, 1–11. Takata, H., Takayasu, K., Hasegawa, S., 2006. Foraminifera in an organic-rich, brackish- Mori, N., Takahashi, T., Yasuda, T., Yanagisawa, H., 2011. Survey of 2011 Tohoku earth- water lagoon, Lake Saroma, Hokkaido, Japan. Journal of Foraminiferal Research 36, quake tsunami inundation and run-up. Geophysical Research Letters 38 http:// 44–60. dx.doi.org/10.1029/2011GL049210. Tamura, T., Masuda, F., 2004. Inner shelf to shoreface depositional sequence in the Sen- Morton, R., Gelfenbaum, G., Jaffe, B., 2007. Physical criteria for distinguishing sandy tsu- dai coastal prism, Pacific coast of northeastern Japan: spatial and temporal growth nami and storm deposits using modern examples. Sedimentary Geology 200, patterns in relation to Holocene relative sea-level change. Journal of Asian Earth 184–207. Sciences 23, 567–576. Nanayama, F., Shigeno, K., 2004. An overview of onshore tsunami deposits in coastal Tamura, T., Masuda, F., 2005. Bed thickness characteristics of inner-shelf storm de- lowland and our sedimentological criteria to recognize them. Memoirs of the posits associated with a transgressive to regressive Holocene wave-dominated Geological Society of Japan 58, 19–33. shelf, Sendai coastal plain, Japan. Sedimentology 52, 1375–1395. Nanayama, F., Shigeno, K., 2006. Inflow and outflow facies from the 1993 tsunami in Tyson, R.V., 1995. Sedimentary Organic Matter: Organic Facies and Palynofacies. Chapman southwest Hokkaido. Sedimentary Geology 187, 139–158. Hall, London. Okahashi, H., Akimoto, K., Mitamura, M., Hirose, K., Yasuhara, M., Yoshikawa, S., 2002. Uchida, J., Fujiwara, O., Hasegawa, S., Kamataki, T., 2010. Sources and depositional pro- Event deposits found in coastal marsh in Osatsu, Toba, central Japan. Chikyu cesses of tsunami deposits: analysis using foraminiferal tests and hydrodynamic Monthly 24, 698–703. verification. Island Arc 19, 427–442. J.E. Pilarczyk et al. / Sedimentary Geology 282 (2012) 78–89 89

Vane, C.H., Drage, T.C., Snape, C.E., 2003. Biodegradation of Oak (Quercus alba) wood Wilson, G.P., Lamb, A.L., Leng, M.J., Gonzalez, S., Huddart, D., 2005. δ13C and C/N as poten- during growth of the shiitake mushroom (lentinula edodes): a molecular approach. tial coastal palaeoenvironmental indicators in the Meresy Estuary, UK. Quaternary Journal of Agricultural and Food Chemistry 51, 947–956. Science Reviews 24, 2015–2029. Vane, C.H., Kim, A.W., McGowan, S., Leng, M.J., Heaton, T.H.E., Kendrick, C.P., Coombs, Yamada, I., Shoji, S., 1981. The Late Holocene felsic tephra in Miyagi Prefecture. Journal P., Yang, H., Swann, G.E.A., 2010. Sedimentary records of sewage pollution using of the Pedological Society of Japan 52, 155–158. faecal markers in contrasting peri-urban shallow lakes. Science of the Total Envi- ronment 409 (2), 345–356.